Perspective | Open Access
Haibin Duan, Xiaobin Xu, "Create Machine Vision Inspired by Eagle Eye", Research, vol. 2022, Article ID 9891728, 5 pages, 2022. https://doi.org/10.34133/2022/9891728
Create Machine Vision Inspired by Eagle Eye
Eagle, a representative species in the raptor world, has the sharpest visual acuity among all animals. The reputation of the “clairvoyance” is employed to describe an eagle. The excellent visual skills of eagles depend on their unique eye structures and special visual principles. The powerful vision perception mechanisms of the eagle bring abundant inspiration for traditional visual applications. Biological eagle eye vision technology provides a creative way to solve visual perception issues of “Knowing What is Where by Seeing.” The theoretical research and practical works of eagle vision would contribute to the development of machine vision, or even artificial intelligence (AI) in the real world. Furthermore, eagle eye vision also provides feasible ideas for the popularization of new concepts in the virtual world in the future.
About 540 million years ago, creatures evolved the eye [1–3]. More than 80% information of the objective world is perceived through the eyes. Since then, the activity mode and survival law of creatures have changed . Eagles have evolved the sharpest vision among all animals in the long process of evolution . The fast and accurate visual information processing abilities of eagles are lost in wonder, which mainly benefits from unique physiological structures and visual processing mechanisms. Using optical devices to imitate the physiological structures of eagle vision is important to improve the performance of hardware devices. Besides, some visual processing procedures of eagles are closely related to computer vision processing technology, especially in high dynamic, high precision, large view field, and high complexity mission environment. More importantly, machine vision is an essential branch of the artificial intelligence (AI) field in the real world. Studying and simulating both the special structures and mechanisms of the eagle eye open up a new train for the development of machine vision. Furthermore, mechanisms of eagle vision can be imported to the virtual world to enhance the security, availability, and experience of the virtual space.
2. Unique Physiological Structure
Eagle vision system is so outstanding that the prey can be locked and tracked outside several kilometers. The phenomenon is inseparable from the unique physiological structure of the eagle eye. Heavily visual-dependent species and highly sensitive organisms have larger eye sizes . Eagle eyes account for 15% of the weight of their head, while that of humans accounts for only 2% . Therefore, the area of an image projected on the retina is increased because of the big tubular eyes of eagle [8, 9]. The photoreceptor cells on the eagle retina can reach up to 1 million/mm2 (that of human is 200 thousand/mm2), so the spatial resolution is higher . Eagle eye’s high spatial resolution can be employed in scenes that pay more attention to detail detection, including intelligent medical treatment, defect detection, and aerial image analysis. Each eye of eagle includes two foveas: median fovea (deep fovea) and lateral fovea (shallow fovea) . Median fovea, used for monocular vision, is more suitable for objects at long distances. The lateral fovea plays a major role in binocular vision for observing objects at close distances. Two foveas cooperate to adapt to the hunting scene . Monocular vision and binocular vision constitute the total view field of eagle . Raptors have different view fields due to the influence of their living environment . The view field of eagle covers 260° and 80° in horizontal direction and vertical direction, respectively . Notably, the large view field structure of eagle provides a new solution for designing a panoramic camera. Based on the above features of eagle eye, Deng and Duan invented an integrated variable resolution imaging device .
Eagles have excellent color recognition ability and light adaptability. The main structures related to the above abilities include photoreceptors, pecten, and oil droplets. The photoreceptors on the retina of eagle include cone cells and rod cells . Cone cells, only located in the fovea, are sensitive to strong light and colorful colors. Rod cells, distributed around the fovea, are related to weak light and achromatic color . Oil droplets (red, green, blue, yellow, and colorless) with a high concentration of carotene are distributed at the end of cone cells . The proportion of oil droplets adjusts according to the living environment. Besides, Seifert et al. suggested that the distribution of cone cells is hyperuniformity, even more perfect than it . It may be the reason why birds have excellent vision, especially the eagle.
3. The Impeccable Optic Nerve System
The process of the vision information processing is inseparable from the participation of the optic nerve system of eagle [20–22]. The visual nervous system of eagle mainly contains four pathways: thalamofugal pathway, tectofugal pathway, retinofugal pathway, and accessory optic system . Four pathways are complementary and indispensable. After decades of research on anatomy, biophysics, and neuroethology, it has shown that these capabilities of color discrimination, luminance adaptation, shape recognition, and moving target detection of eagle are invalid without the visual nervous system .
Optic tectum, an important part of the tectofugal pathway or midbrain, is the visual center of eagle for fusing and transmitting vision information. Optic tectum and nucleus isthmi together comprise competition and selection branch for input visual stimulations in the tectofugal pathway . Visual attention mechanism, lateral inhibition mechanism, and winner-take-all mechanism occurred in optic tectum-nucleus isthmi pathway  are investigated. These mechanisms of eagle vision have been applied in target detection, target tracking, contour extraction, autonomous air refueling, and autonomous carrier landing.
Moving object detection is also closely related to the optic nerve system of eagle. Moving target detection mainly focuses on two aspects: fault-tolerant ability and optical flow sensitivity. Parallax sensitive cells, accommodating a certain range of visual noise, exist in the visual cortex of eagle. Eagle is expert in observing and estimating the movement of prey through optical flow. Therefore, eagle can accurately find, track, and capture prey in complex and variable scenes.
4. Future Eagle Vision Applications
Vision technology is a critical component of intelligent machine or AI. Machines with keen vision system can commendably accomplish complex, dangerous, and tedious tasks in various environments. In view of the above requirements, machine vision puts forward higher standards for accuracy and adaptability. The exploration of eagle vision mechanisms can provide an effective development route for modern machine vision (Figure 1). The typical contents of the route mainly include physiological structure, biological mechanism, computational models, hardware device and microchips, and practical applications. Moreover, eagle vision can also bring extraordinary experience for human society with the digital age coming.
4.1. Investigation of the Biological Mechanism
Some mechanisms based on eagle vision were established in the past years, such as contrast sensitivity mechanism, color antagonism mechanism, visual attention mechanism, and lateral inhibition mechanism . The above mechanisms of eagle vision provide feasible solutions for machine vision. As the foundation of eagle vision technology, the physiological structure and functional characteristics of eagle eye are expected to be further investigated. Furthermore, the function and mechanism of different information processing units for biological eagle eye vision system should be analyzed to improve the detection accuracy and speed of machine vision in complex and dynamic mission circumstances.
The exploration of undiscovered eagle vision mechanism has a significant effect for developing the biological eagle vision mechanism continuously and deeply. The visual signal processing of eagle eye requires the joint participation of the eye, optic nerve, and brain. For this reason, the only way which must be passed is to establish a comprehensive and integrated eagle eye-nerve-brain visual information processing system model for the research of eagle vision. Therefore, on the premise of complying with the restrictions of animal protection policies, it is necessary to continue the relevant behavioral experiments of eagle eye with the form of minimally invasive by cooperation in various fields.
4.2. Computational Physiological Model
The computational physiological model is highly expected to be established using the information processing mechanism of eagle vision . The typical models include the retinal structure model, the oil adaptive distribution model, and the receptive field model . Using these models to simulate the information processing mechanisms of eagle vision is critically beneficial to applying these mechanisms in improving machine vision. Additionally, as is well-known, the visual pathway of eagle midbrain is an important part in the eagle vision system. Therefore, through analyzing the relationship between the brain pathway and the visual nucleus, the eagle vision-brain mechanism model could be fundamentally built, which can play a significant role in perfecting the eagle eye mechanism model.
4.3. Multidisciplinary Study in Eagle Eye-Brain-Cognition-Behavior Technology
The development of biological eagle vision is still in its infancy. The existing information processing mechanisms of eagle vision system are mainly skilled in long-distance small target searching, large view field target tracking, and strong night vision detection. Eagle eye-brain-cognition-behavior mechanism is the complement and improvement for current mechanisms . It requires the joint participation of multidisciplinary, including zoology, biology, biophysics, biochemistry, neurology, ethology, optoelectronics, anatomy, physics, electronics, cybernetics, and engineering . With the rapid development of AI, machine vision based on biological eagle eye will have characteristics of autonomous learning, data sharing, and human-computer cooperation. Biological eagle vision has important enlightening significance for the diversified development of modern machine vision.
4.4. Eagle Vision-Based Devices and Microchips
Biological eagle eye can be simulated from physiological structures or functions. Some sample hardware devices have been invented to simulate the large view field characteristic and variable resolution characteristic of eagle [30–33], and some of the devices are applied in autonomous aerial refueling, robot swarm formation, missile remote guidance, and smart city. However, the combination of eye-brain-cognition-behavior mechanism and microchips is still on the road. Miniaturization is the premise for improving the utilization rate of biological eagle eye devices. Microchips that load with eagle vision mechanisms are crucial to the development of the machine information industry. Besides, the bionic inventions inspired by eagle’s claws and wings are important for farmland protection, camouflage reconnaissance, search, and rescue, which is also a promising way for biological eagle eye device and microchip design.
4.5. Cross-Domain Applications Based on Eagle Vision in the Real World
In the real world, highly intelligent machines can be regarded as the most ideal tools for enriching human behaviors. Generally, high-resolution and precise vision is one of the important evaluation indexes of machine intelligence . Biological eagle vision is an important technical means to promote the development of machine vision or AI. The technology can be applied for smart transportation, smart agriculture, smart medical, smart catering, geological exploration, and other civil and military industries. Cross-domain  applications mainly focus on the combination ability and cooperation ability between different spaces and individuals. The performance of the whole system will be significantly improved when each machine in different domain has “smart eagle eye.” Particularly, eagle vision may also have important application value on the land, in air, in ocean and deep sea, and also in space integrative tridimensional military field, all of them are challenging areas for machine vision technologies.
4.6. Eagle Vision-Based Applications in the Virtual World
Recently, metaverse  is a hot point in virtual reality, which is a 1: n projection from the real world to the virtual world. Metaverse emphasizes various experiences and interactions of virtual scenes and describes the virtual world through real logic or imaginary logic [37, 38]. Eagle is not only the king of the sky but also the king of vision. If human or machine has a pair of “eagle eyes,” it can capture not only the details of the scenery in front of it, but also the scenery thousands of miles away with wide vision field. Moreover, it will experience the beauty of four-color space. Objects in the scene can be seen even in the weak light environment. Then, it will experience extraordinary scenes without visual blind spots in the virtual world.
The excellent vision of eagle has attracted many biologists for a long time. And many scientists and researchers from other fields also paid abundant attention to this cross-domain field. Biological eagle vision technology provides an effective approach for the enable the modern machine vision or unfolding AI with a more bright and smart “eye.”
Conflicts of Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
All authors contributed to the paper discussion and editing.
This study was supported by the Science and Technology Innovation 2030-Key Project of “New Generation Artificial Intelligence” (2018AAA0102303), National Natural Science Foundation of China (U1913602, U20B2071, 91948204, T2121003, and U19B2033), and Academic Excellence Foundation of BUAA for PhD Students under grant #202115.
- J. R. Paterson, D. C. García-Bellido, M. S. Y. Lee, G. A. Brock, J. B. Jago, and G. D. Edgecombe, “Acute vision in the giant Cambrian predator Anomalocaris and the origin of compound eyes,” Nature, vol. 480, no. 7376, pp. 237–240, 2011.
- J. Paterson, G. D. Edgecombe, and D. C. García-Bellido, “Disparate compound eyes of Cambrian radiodonts reveal their developmental growth mode and diverse visual ecology,” Science Advances, vol. 6, no. 49, article eabc6721, 2020.
- H. Zeng, F. Zhao, K. Niu, M. Zhu, and D. Huang, “An early Cambrian euarthropod with radiodont-like raptorial appendages,” Nature, vol. 588, no. 7836, pp. 101–105, 2020.
- B. Schoenemann, H. Pärnaste, and E. N. Clarkson, “Structure and function of a compound eye, more than half a billion years old,” Proceedings of the National Academy of Sciences of the United States of America, vol. 114, no. 51, pp. 13489–13494, 2017.
- T. J. Lisney, K. Stecyk, J. Kolominsky, G. R. Graves, D. R. Wylie, and A. N. Iwaniuk, “Comparison of eye morphology and retinal topography in two species of new world vultures (Aves: Cathartidae),” Anatomical Record, vol. 296, no. 12, pp. 1954–1970, 2013.
- S. Potier, M. Mitkus, and A. Kelber, “Visual adaptations of diurnal and nocturnal raptors,” Seminars in Cell and Developmental Biology, vol. 106, pp. 116–126, 2020.
- R. Shlaer, “An eagle's eye: quality of the retinal image,” Science, vol. 176, no. 4037, pp. 920–922, 1972.
- S. Potier, F. Bonadonna, A. Kelber, and O. Duriez, “Visual acuity in an opportunistic raptor, the chimango caracara (Milvago chimango),” Physiology & Behavior, vol. 157, pp. 125–128, 2016.
- G. R. Martin, “Visual fields and their functions in birds,” Journal of Ornithology, vol. 148, no. S2, pp. 547–562, 2007.
- W. M. Harmening, P. Nikolay, J. Orlowski, and H. Wagner, “Spatial contrast sensitivity and grating acuity of barn owls,” Journal of Vision, vol. 9, no. 7, pp. 1–12, 2009.
- A. W. Snyder and W. H. Miller, “Telephoto lens system of falconiform eyes,” Nature, vol. 275, no. 5676, pp. 127–129, 1978.
- G. R. Martin, “The subtlety of simple eyes: the tuning of visual fields to perceptual challenges in birds,” Philosophical Transactions of the Royal Society B Biological Sciences, vol. 369, no. 1636, article 20130040, 2014.
- S. P. Mysore and N. B. Kothari, “Mechanisms of competitive selection: a canonical neural circuit framework,” Elife, vol. 9, no. 9, article e51473, 2020.
- G. R. Martin, S. J. Portugal, and C. P. Murn, “Visual fields, foraging and collision vulnerability in Gyps vultures,” Ibis, vol. 154, no. 3, pp. 626–631, 2012.
- Y. Deng and H. Duan, “Biological eagle-eye based visual platform for target detection,” IEEE Transactions on Aerospace and Electronic Systems, vol. 54, no. 6, pp. 3125–3136, 2018.
- R. Fox, S. Lehmkuhle, and D. Westendorf, “Falcon visual acuity,” Science, vol. 192, no. 4236, pp. 263–265, 1976.
- J. Hirsch, “Falcon visual sensitivity to grating contrast,” Nature, vol. 300, no. 5887, pp. 57-58, 1982.
- A. Guzman-Pando and M. I. Chacon-Murguia, “DeepFoveaNet: deep fovea eagle-eye bioinspired model to detect moving objects,” IEEE Transactions on Image Processing, vol. 30, pp. 7090–7100, 2021.
- M. Seifert, T. Baden, and D. Osorio, “The retinal basis of vision in chicken,” Seminars in Cell and Developmental Biology, vol. 106, pp. 106–115, 2020.
- P. H. Schiller, “Parallel information processing channels created in the retina,” PNAS, vol. 107, no. 40, pp. 17087–17094, 2010.
- H. Duan, X. Xu, Y. Deng, and Z. Zeng, “Unmanned aerial vehicle recognition of maritime small-target based on biological eagle-eye vision adaptation mechanism,” IEEE Transactions on Aerospace and Electronic Systems, vol. 57, no. 5, pp. 3368–3382, 2021.
- X. Li, H. Duan, J. Li, Y. Deng, and F. Y. Wang, “Biological eagle eye-based method for change detection in water scenes,” Pattern Recognition, vol. 122, article 108203, 2022.
- A. Querubin, H. R. Lee, J. M. Provis, and K. M. B. O'Brien, “Photoreceptor and ganglion cell topographies correlate with in formation convergence and high acuity regions in the adult pigeon (Columba livia) retina,” Journal of Comparative Neurology, vol. 517, no. 5, pp. 711–722, 2009.
- A. P. Nguyen, M. L. Spetch, N. A. Crowder, I. R. Winship, P. L. Hurd, and D. R. Wylie, “A dissociation of motion and spatial-pattern vision in the avian telencephalon: implications for the evolution of “visual streams”,” Journal of Neuroscience, vol. 24, no. 21, pp. 4962–4970, 2004.
- V. A. Tucker, “Gliding flight: drag and torque of a hawk and a falcon with straight and turned heads, and a lower value for the parasite drag coefficient,” Journal of Experimental Biology, vol. 203, no. 24, pp. 3733–3744, 2000.
- H. Duan, Y. Deng, and X. Wang, Biological Eagle-Eye Vision and Its Applications, Science Press, Beijing, 2021.
- Y. Sun, Y. Deng, H. Duan, and X. Xu, “Bionic visual close-range navigation control system for the docking stage of probe-and-drogue autonomous aerial refueling,” Aerospace Science and Technology, vol. 91, pp. 136–149, 2019.
- M. P. Jones, K. E. Pierce, and D. Ward, “Avian vision: a review of form and function with special consideration to birds of prey,” Journal of Exotic Pet Medicine, vol. 16, no. 2, pp. 69–87, 2007.
- M. Zirnsak, N. A. Steinmetz, B. Noudoost, K. Z. Xu, and T. Moore, “Visual space is compressed in prefrontal cortex before eye movements,” Nature, vol. 507, no. 7493, pp. 504–507, 2014.
- X. Du, J. Chang, Y. Zhang et al., “Design of a dynamic dual-foveated imaging system,” Optics Express, vol. 23, no. 20, pp. 26032–26040, 2015.
- A. D. Long, R. M. Narayanan, T. J. Kane, T. F. Rice, and M. J. Tauber, “Analysis and implementation of the foveated vision of the raptor eye,” in Proceedings of SPIE Conference on Image Sensing Technologies, Baltimore, MD. USA, 2016.
- P. B. Melnyk and R. A. Messner, “Biologically motivated composite image sensor for deep field target tracking,” in Proceedings of SPIE-the International Society for Optical Engineering, USA, 2007.
- L. Lin, B. Ramesh, and C. Xiang, “Biologically inspired composite vision system for multiple depth-of-field vehicle tracking and speed detection,” in Proceedings of Asian Conference on Computer Vision, pp. 473–486, Berlin, Germany, 2015.
- M. Li, L. Fu, and X. Wang, “A cross-domain augmentation-based AI learning framework for in-network gesture recognition,” IEEE Network, vol. 35, no. 5, pp. 90–97, 2021.
- Y. Xu, J. Hu, K. Wattanachote, K. Zeng, and Y. Y. Gong, “Sketch-based shape retrieval via best view selection and a cross-domain similarity measure,” IEEE Transactions on Multimedia, vol. 22, no. 11, pp. 2950–2962, 2020.
- M. A. I. Mozumder, M. M. Sheeraz, A. Athar, S. Aich, and H. C. Kim, “Overview: technology roadmap of the future trend of metaverse based on IoT, blockchain, AI technique, and medical domain metaverse activity,” in 2022 24th International Conference on Advanced Communication Technology (ICACT), pp. 13–16, Korea, 2022.
- M. Abrash, “Creating the future: augmented reality, the next human-machine interface,” in Proceedings of IEEE International Electron Devices Meeting, pp. 1.2.1–1.2.11, San Francisco, CA, United States, 2021.
- A. A. Sheikh and M. Junaid, “Machine learning inspired vision-based drowsiness detection using eye and body motion features,” in Proceedings of 13th International Conference on Information and Communication Technology and System, Virtual, Online, pp. 146–150, Indonesia, 2021.
Copyright © 2022 Haibin Duan and Xiaobin Xu. Exclusive Licensee Science and Technology Review Publishing House. Distributed under a Creative Commons Attribution License (CC BY 4.0).